Our project aims to detect whether sentence pairs have the same meaning or not. Our reference dataset is the PAWS, which is Paraphrase Adversaries from Word Scrambling. The second sentence of each pair is generated by scrambling the first one with a certain criteria.
The project does not have loaded models in it, but train different models for every time it runs. Also, it was developed in a Google Colab environment, so its working is secured in such environment. If one wants to run this in a local environment, they should make sure that each library used in our project that is already installed in Google Colab is installed as well.
- Caviglia Matteo
- Sofia Gaia Formenti